BBY Fehler 17584 Lambda-Sonde nach Katalysator | Posts 32+

Hi Jan,
1.5 Ohm would be a bit too low for a line. How much did you actually measure?
hg
Herbert

Hi,

I measured just under 1.5 Ohms. However, the problem was less due to the wiring itself and more due to the setup with various connected/clamped cables (it had to reach from under the car to the MSG connector), and unsuitable contacts (thin pin-needle contacts).

Since I also found this a bit lacking, I ended up holding the ends of my "measuring wire" together at the end, using the same setup. This resulted in approximately 1.2 Ohms. The cause therefore really lies there.

RLF says <1.5 Ohm - that's why I was also satisfied with it. Whether it's 0.3 or 0.1 Ohms, it doesn't really matter for the problem.

The three wires that I measured (the two signal wires, as well as the switching side of the probe heater) all had exactly the same resistance when measured under the same conditions.

Therefore, I refrained from optimizing the measurement setup, because if there are any weaknesses in the cable, they will certainly not affect all three wires in exactly the same way, especially since the heating wire is being monitored (and that works, the tests without a probe threw up corresponding errors).

The two sonde leads have an infinite resistance between them, according to the meter. So, wenn also das Steuergerät eine Referenzspannung von 0,4-0,5V ausgibt, sollte diese am anderen Ende (ohne Last angeschlossen) auch ankommen, selbst wenn es einen tatsächlichen Leitungswiderstand von 1,5 Ohm gäbe.

The measuring device is a Peaktech 3315U or 3415U, I need to double-check at home. In any case, it's definitely not a cheap hardware store item.

What I unfortunately can't easily test is whether the control unit outputs the reference voltage even without any connected devices, because to do this, you would have to operate it on a workbench in such a way that it believes a car is connected and that the ignition is switched on. And even then, you wouldn't know without a comparison device whether it is actually outputting the reference voltage in the simulated environment or not. If someone knows exactly how to make an existing control unit output the reference voltage for the post-catheter sonde to pins 68 and 69, then I would be happy to test it.

Best regards,

Jan

Hi,

and one more thing that just occurred to me: What neither my measurement method nor the RLF covers is a short circuit (or a contact resistance) to another line in the cable harness, which has a high impedance towards the car and is connected to some potential by the MSG. That would not be possible either when measuring the contacts of the lambda sensor against each other or against the ground, if the MSG was disconnected.

If a standalone measurement of the MSG is not feasible or reliable, it is probably easiest to obtain an MSG from a book or similar and try it out. For the reference voltage measurement, the WFS does not even need to be calibrated. Then you could measure with a probe - or even simpler, look at the MWB, as the current device always has exactly the same thing inside, which I measure at the probe connector. Can I access the MWB of a MSG with an unlearned WFS using VCDS?

Best regards,

Jan

Hello Jan,

Although I don't have more detailed information, I know that with the 1.4 petrol engines, problems were sometimes caused by the original crimp contacts at the connection to the sensor. In these cases, the original contacts were replaced with new ones that were coated with gold. A colleague had two (!) such Beetles running in parallel, and both were permanently cured by VW through the replacement of the contacts.

Hi,

Yep, I know that... However, the Polo already has contacts with gold plating at that location. However, the contacts of the sonde heater are normal, therefore the difference is immediately apparent.

However, this is irrelevant for the current problem: When measuring the ground voltage/reference voltage, you measure it with the probe disconnected, and you should be able to measure something (or nothing) regardless of the material of the contacts (or if the contacts are completely worn out).

Best regards,

Jan

70mV is practically nothing... if you touch the contact a little or have an inductive/capacitive coupling, especially with a multimeter with 3.5 digits, then something like that can sometimes be displayed.

Have you ever connected the cable all the way to the MSG? Make a loop on the plug and find the contacts on the MSG plug.

Greetings, from Styria.

Hi,

siehe a few posts above. I have disconnected the probe and MSG, and then measured for continuity (the pins are known... 68 and 69 are signal, 63 is the switched side of the probe heater). All the wires have continuity, there is no resistance to measure between them. Also, no resistance can be measured to ground (except for the probe heater's ground wire). From this, I conclude that the cable is okay. The only error that my measurement method would not be able to detect is a short circuit to another wire in the engine cable harness, which is pulled up to this level on the MSG side.

Quote:

If you're fiddling around with the contacts or there's an inductive/capacitive coupling, especially with a 3.5-digit meter, then you might sometimes see something like that.

The value is very stable and reproducible. Similarly, it changes reproducibly between motor standstill (around 60 mV) and motor running (around 70 mV). Tweaking around won't help; the measuring device has full four decimal places and is not a cheap hardware device. The control unit itself measures (with only a few mV difference) the same value and displays it on the MWB 36. (I ran the engine for a while without the after-catalytic converter sensor and monitored the MWB 36, measuring it repeatedly - the result was always the same).

The next logical test would be to measure the signals at the pins of the MSG, without the cable harness attached. "However, I'm still not clear on how to achieve this so that the MSG actually outputs the reference voltage (or at least should), while minimizing the amount of connected peripherals."

Does anyone happen to have a Magneti-Marelli ECU (this is a 4MV, but the exact type probably doesn't matter for testing) running in a bench diagnostic setup and can measure between the pins of the after-fire sensor when "ignition on"? For the 036 906 034 DE, these are pins 68 and 69. If so, it would also be interesting to know what minimal additional components need to be connected (ideally just the power supply).

Best regards,

Jan

...I missed the MSG part, sorry. I thought you were only measuring at the disconnected poles of the probe. But around 70mV sounded like an open end to me, meaning a hum from somewhere. So, mass errors and so on...

Greetings, from Styria.

Hi,

Another update: Engine control unit removed, powered up, and then measured. Still only 60 mV at the lambda sensor connections. Under these circumstances, I would have expected either 0 volts (because the car is not connected or not connected enough) or the 0.45 volts of the reference voltage. However, this is exactly what also happened in the car, so I conclude: The wiring is also okay, as is the lambda sensor. The engine control unit has a defect in the relevant part of the circuit.

Best regards,

Jan

Hi,

Here's the translation:

The RLF is incorrect. The base voltage was not correct, so I purchased a used second motor control unit. Result: still only 60 mV. Either the RLF is incorrect, or both MSG (which are only slightly more than a year apart in production date) have the same error.

So... the next attempt was to see what the MSG (Measurement System) on the table said when presented with a voltage of 0-1V with an impedance of approximately 10 kOhm (similar to a lambda probe, simulated with a lab power supply + resistor). It displayed this voltage without any issues on the MWB 36 and appears to have no fault in the form of an input that is too low in resistance. Therefore, it should also work with a real lambda sensor, and therefore it is probably more likely that the sensor itself is the problem (as the RLF had indicated), because it is either not reaching the required voltage or has too high an impedance, which means it is probably broken.

Today, I installed a new replica probe... everything is perfect. The values are correct (except for the initial ground voltage, which is still 60 mV - that's probably normal for a cold probe), and there are no errors when driving (the probe behaves exactly as it should as a jump probe). The readiness can be set, so everything is now in perfect order.

The RLF has therefore led us astray... the whole thing could have been quick and easy.

But perhaps this information will be helpful to someone else with the same problem.

Best regards,

Jan